NREL/TP-262-4678 • UC Category: 331 • DE92016431

Environmental, th, and Safety Issues of Sodi-...... Batteries for Electric and rid Vehicles

V()lume 1: Cell. attery Safety

J. M. Obi

National Renewable Energy Laboratory (formerly the Solar Energy Research Institute) 1617 Cole Boulevard Golden, Colorado 80401-3393 A Division of Midwest Research Institute

· Operated for the U.S. Department of Energy under Contract No. DE-AC02-83CH10093

Prepared under Task No. AS015440

September 1992 l l.

r l

L

NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor any agency thereof. nor any of their employees. makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, com­ pleteness, or usefulness of any information, apparatus, product, or process disclosed. or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name. trademark, manufacturer, or otherwise does not necessarily con­ stitute or imply its endorsement, recommendation. or favoring by the United States government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.

Printed in the United States of America Available from: National Technical Information Service U.S. Department of Commerce 5285 Port Royal Road Springfield, VA 22161 'c L

Price: Microfiche A01 Printed Copy A03

Codes are used for pricing all publications. The code is determined by the number of pages in the publication. Information pertaining to the pricing codes can be found in the current issue of the following publications which are generally available in most libraries: Energy Research Abstracts (ERA); Govern­ ment Reports Announcements and Index ( GRA and I); Scientific and Technical Abstract Reports (STAR); and publication NTIS-PR-360 available from NTIS at the above address. TP-4678 ·

I Preface 1 This reportis thefirst of four volumesthat identify and assess the environmental, health,and safety issues .. involved in using sodium-sulfur (Na/S) battery technologyas the energy source in electric and hybrid vehicles thatmay affectthe commercialization of Na/S batteries. This andthe other reportson recycling, shipping, and vehicle safety are intended to help the Electric andHybrid Propulsion Division of the I Office of Transportation Technologies in theU.S. Department of Energy (DOFJEHP)determine the direction of its research, development, and demonstration (RD&D) program for Na/S battery technology.1 r The reports review thestatus of Na/Sbattery RD&D andidentify potential hazards and risks thatmay require additional reseaiCh or thatmay affect the design and use of Na/S batteries.

These reports were prepared by the Analytic Studies Division of theNational Renewable Energy Laboratory andare one part of DOFJEHP's RD&D program to work with industry to commercialize Na/S batteries. For example, data andinformation obtained through these reportswill assist the DOFJEHP implement recommendations made by participants at government-industry meetings on sodium-beta batteries sponsored by the DOFJEHP [1]. The reports may also assist the DOFJEHPand the Ad Hoc Electric VehicleBattery Readiness WorkingGroup coordinate the RD&D needed to commercialize Na/S and sodium metal chloride batterytechnologies. 2

For these reports, it is important to define hazard and risk. A hazard is a source of risk, a substance or

action that can causeharm . Risk, onthe otherhand, is the possibility of suffering harm from a hazard [2]. While the chemical and thermal hazards of elemental sodium are substantial, the risks involved in using sodium in a battery can be minimized through careful design, engineering, andtesting. These reports on Na/S battery technologydo not constitute a formalrisk analysis, which usually includesestimates of the amounts, frequencies, andlocations of the releaseof hazardous materials; the durationof exposures to these agents; estimates of the percentage of the population exposed and of dosage-response relationships; and a quantitative estimate of risk [2]. These reports provide a qualitative analysis of hazards andrisks that must be addressed before Na/Sbatteries can be deployed on a commercial scale. These reports are intended to help DOFJEHP set management priorities for the RD&D of Na/S battery technology by identifying potential hazards andrisks, by reviewing RD&D in progress to address these hazards and risks, and by recommendingRD&D needed to help minimizethese hazards and risks.

This volume covers cell design andengineering as the basis of safety for Na/S batteries and describes and assesses the potential chemical, electrical, and thermalhazards and risks of Na/S cells and batteries as well as the RD&D performed, under way, or needed to address these hazards and risks. The repOrt is based on a review of the literatureand on discussions with experts at DOE,national laboratories and

1 Theseassessments are concerned with Na/S batteries usedas traction batteries for electric and hybrid vehicles and are notconcerned with stationary energy storage.

2 The Ad Hoc Electric Vehicle Battery Readiness Working Group consists of leading scientists and program managers from government agencies, battery developers, automobile manufacturers, and the chemical processing industry.

iii TP-4678 agencies, universities, andprivate industry. Subsequent volumes will addressenvironmenta l, health, and safety issues involved in shipping cells and batteries, using batteries to propel electric vehicles, and recycling and disposing of spent batteries.

f '·· The remainder of this volume is divided into two major sections on safety atthe cell and batterylevel s. The section on Na/Scells describes major components andpotential failure modes, design, life testing and failure testing, thermalcycling, and the safety status of Na/S cells. The section on batteries describes battery design, testing, and safety status. Additional EH&S information on Na/S batteriesis provided in the appendices. i

I am indebted to many people who helpedme obtain information and reviewed drafts of this volume. I t am especially indebted to J. Braithwaite of Sandia National Laboratories who helped me understand some of the technical intricacies of Na/S battery technology and who patiently reviewed several drafts of this volume. I would also like to thankR. MacDowall of IdahoNational Engineering Laboratory, J. Smaga of Argonne National Laboratory, E. Dowgiallo, Jr., andK. Heitner ofOOFJEHP, H. Haskins of Ford Motor Company and the USABC, and R. Swaroop of EPRI for reviewing a draftof this volume and offering helpful technical suggestions on the text. I am also indebted to J. Rasmussen of Beta Power, Inc., for providing me technical insights into Na/S batteries, especially those concerning the beta"-alumina electrolyte. lalso thank W. Fischer of ABB and P. Binden of CSPL for sending me information and publications on safety issues and M. Brada(then) of the Colorado School of Mines for assisting me in research. Finally, I thankD. O'Hara ofDOEIEHP for his leadership and support in directing and sponsoring theEH&S program andthese assessments.

iv TP-4678

Table of Contents

Page No.

Introduction 1

Intrinsic Material Hazards ...... 3

Na/S Cell Components and Failure Modes 3

Background •...... 3 Beta"-Alumina Electrolyte 5

Seals ...... 6 Alpha-Alumina-Beta "-Alumina Glass 6

Metal-to-Alpha-Alumina •..... 6 Cell Container-Sulfur Electrode 7

Na/S Cell Design ...... 7

Cell Safety Testing ...... 9

Life Testing 9 Failure Testing 11 Thermal (Freeze-Thaw) Cycling 13

Safety Sta.tus of Na/S Cells ...... 14

Battery Safety • • . . . • 14

Battery Design 14 Battery Failure Testing 16 Battery Life-Testing 17

Safety Status of Na/S Batteries 20

Conclusions and Recommendations 21

References ...... 23

Appendix A- Chemical Hazards of Sodium-Sulfur Batteries ...... A-1

Appendix B - Toxicities and Reactivities of Chemical Elements and Compounds Potentially Associated with the use

of Na/S Batteries ...... B-1

v TP-4678 f I; · List of Tables '

Page No. f L Table 1 - Hazardous Chemical Species Associated with r a Sodium-Sulfur Battery . . • • • • . . • • . . • • . • . . . • • ...... • 4

Table 2- Cell Component Stability • ...... • ...... 12 r ; Table 3- Program for Safety Tests on Sodium-Sulfur Batteries L. Part 1: Normal Use ...... 18

Table 4- Program for Safety Tests on Sodium-Sulfur Batteries

Part 2: Safety During/After Accidents .•...... •...... 19

List of Figures

Page No.

Figure 1- The N �SIS phase diagram . . • ...... 2

Figure 2- The ABB cell design ...... • • . • • . . . . . • ...... 8

Figure 3- The CSPL cell design . . • ...... 10

L vi TP-4678

Introduction

A Na/S battery consists of three major subsystems: a large number of mechanically andelectrically interconnected cells, a thermal enclosure to maintain an operating temperature range of 300°-350°C, and a thermal management system for initial heating of the battery and for removing waste heat [3]. System components peripheral to the battery itself, such as the battery controller, affect battery safety but will not be considered in this report.

The major components of the Na/S cell area solid ceramic electrolyte of beta" -alumina, liquid electrodes of sodium and sulfur, and a container [4]. The electrolyte is usually sealed to an alpha-alumina component that provides aninsulating surface for the seals on the metal current-collectors. During most of the dischargetwo liquid phases exist (Figure l ), and the extent of the cell reactionposs ible varies according to the temperature of operation [5]. The theoretical specific energy normallyquoted for the Na/S cell, 760Whlkg, is calculated for the reaction proceeding to an average compositionof Na2S3 :

2Na+ 3 S = Na�3

The electromotive force (EMF) of the cell at the normal operating temperature of 350°C is 2.076 V , which declines to 1.74 Vat a polysulfide composition corresponding to Na2S3 [6].

Development of the Na/Scell, first announced by the FordMotor Company in 1966, was made possible by the discovery ofthe high ionic conductivity of sodium ions in beta-alumina ceramic [6]. In the intervening 25 years,Na/S cells have undergone continual development, and the twoprincipal manufacturers of Na/S batteries fo r electric vehicles, Asea BrownBoveri (ABB) and Chloride Silent Power Limited (CSPL) have made many changes to both cell and batterydesign to improve safety. This reportdescribes the evolution of cell designs at CSPL and ABB to address safety issues and identifiescell design issues requiring additional RD&D. The report also evaluates safety issues involved in interconnecting cells to form modules and batteries.

Although electrical integrity at the battery level is critical and electrical hazards associated with a Na/S (or any high-energy) battery must be acknowledged, the key (and the emphasis of this assessment) for the safe operation of Na/S batteries resides in the chemical, thermal, and structural integrity of the individual cells that make up the battery. Due to the large number of cells, the temperature-dependent nature of the chemical reactions, and the potential for cell-to-cell interactions, however, safety testing at the cell level cannot always be extrapolated to the battery level. In this report, a clear distinction is maintained between cell and battery safety as well asbetween cell and battery testing. A further distinction must be made between safety under normal operating conditions and safety under catastrophic events such as collisions. The latter is also discussed in the sections on failure testing of cells and batteries andin the volume on vehicle safety issues.

l TP-4678

i !_

600 2 Na:4·67S

500

I Vapor+ liquid I 6 400 I 2...... f · Cl) ..... I Single phase I Two liquids ::J \_ co I Na2Sx S + Na2S5 ..... I Cl) 0. 1------t------I Operating f E 2850 I � 300 Na2S2 Na2s3 temperature I I + I liquid �---- 1 200 I Na2S5 I liquid sulfur I I

60 70 80 90 100 Sulfur (wt %)

Figure 1. The NazSIS phase diagram Source: Adapted from Ref. 5.

2 TP-4678

Intrinsic Material Hazards

A battery by definition is a source of direct-current voltageand involves construction of an electrochemical device to allow controlled interaction of reactivematerials. The use of reactive materials is, also by definition, a hazardous enterprise. Inthe caseof Na/S batteries, the hazards of using elemental sodiumand sulfur, which arewell knownas reactive materials, areaccentuated by the need to maintain them in a molten state (about 350°C) separatedby a thin ceramic electrolyte. The chemical and thennal hazards due to primaryreactions of materials used in the Na/S batteryare arizedsumm in Table 1 and AppendixA. Again, it should bestressed that hazards, not risks, arediscussed in this section. Further analysis is requiredto derive a quantitativeestimate of riskfrom a particularhazar d For example, chromium sulfide, a potentialproduct from corrosion of the cellcontainer, is toxic, but its verylow concentrationand insolubility substantially removes the riskof injury from this particularhazard.

Hazards arenot only associated with chemical reactionsthat occur within the cell, as in the example of chromium sulfide, but also when contaminants or outside agents areinadvertently introduced, for example, as a result of an accident or collision. Hazards associated with sodiuminclude its reactionwith water. This reactionis highly exothennic, usually explosive, andproduces hydrogengas and sodium hydroxide [6]. These reaction products can also be hazards. Hydrogen in the presence of oxygen is highly explosive, and sodium hydroxide is poisonous upon ingestion and can cause severe chemical burns [6]. The direct reaction of molten sodium and molten sulfur is a firehazard and produces sodium sulfide. Sodium reacts exothermically with dozens of other elementsand compounds. Itis spontaneously flammable when heated in air and will emit sodiummonoxide, a toxic gas [6]. If sodium bums on concrete, there could be spalling of concrete and splattering of metallic sodium, concrete, and caustic chemical compounds. Inthe presence of water, sodiumsulfides can produce hydrogen sulfide, an explosive and poisonous gas.

Sulfur is not a8 chemically reactive as sodium, but there aresignificant hazards associated withits use. Sulfuris flammable andproduces sulfur dioxide, a toxic gas and a tumorigen, when burned [6]. The reaction of sodiumand sulfur produces sodium polysulfides, which are highly corrosive and can corrode the metallic housing of thecell to fonn chromium sulfide described above. The elements andcompounds discussed in this and the previous paragraphare the principal sourcesof the chemical, physical, and thennal hazards of reactants and reaction products possible from the use or abuse of Na/S cells. The toxicities of these elements andcompounds aregiven in Appendix B.

Na/S Cell Components and Failure Modes

Background

The safety of Na/S batteries depends on how well hazards, which areintrinsic in the reactivityof the materials used in the Na/S cell, areaccounted for in the design of cells and batteries. A typical SO-kWh Na/S battery to power an electric vehicle contains about 40 kg of molten sodium and 60kg of molten sulfur at approximately 350°C. The key to safe perfonnance under nonnal operating conditions is to minimize the occurrence of large thennal excursions by: (1) not to allowing significant quantities of

3 TP -4678

Table 1. Hazardous Chemical Species Associated with a Sodium-sulfur Battery

Species Safety Concern

r l Liquid sodium Hot, toxica , reactive,bums inair, reactsviolently with liquid sulfur, reactsexplosively with water

Liquid sulfur Hot, bums in air,reacts violently with liquid sodium, contaminated with sodium sulfideand sodium polysulfide

Sodium polysulfideb Toxic, can explode on percussion or rapid heating

Sodium oxide Toxic, damageshuman skin andmucous membranes

Sodium hydroxide Toxic, damages human skin andmucous membranes

Sulfur dioxide Toxic, heavier than air

Chromic sulfidec Toxic, not appreciably soluble in water

( Hydrogen sulfide Toxic, lethal in low dosage, explosive J· 'l

• "Toxic" indicates a toxicity rating of 3 as defined inN. I. Sax, Dangerous Properties of Industrial Materials, 6th Edition, Van Nostrand Reinhold, 1984.

l b The sulfur phasecontains varioussulfur polymers, usually designated as polysulfur.Polysulfur reacts with sodium to form sodiumpolysulfide. c Chromium ion may beformed in the sulfur electrode from corrosion of chromium.

4 TP-4678

sodium and sulfur to interact, (2) containing all molten reactants and products, and (3) preventing large­ scale electrical short-circuiting. The approach taken to date is to subdivide the sodium and sulfur into many small, sealed cell packages [5]. Safety of the Na/S battery then becomes dependent, to a large part, on the safety of each cell.

Cell safetyhas beenan important part of Na/S research since the first paper on cell safety was published in 1975 by Hames andTilley who defmed a safe cell as a "cell which, under all but the most improbable conditions, did not allow its reactive contents to escape in the event of cell failure or accident" [5]. Tilley defines three general principles of cell safety [5]:

1. Minimize the quantity of sodium immediately available for reaction after beta"- alumina failure 2. Separate the bulk of cell reactants and restrict the flow of bulk sodium to the reaction site, 3. Protect the outer cell case against corrosion by sodium polysulftdes at elevated temperatures (for a sodium-core cell).

Beta11·Aiumina Electrolyte

The most hazardous situation at the cell level is created by the direct reaction of sodium and sulfur, during which temperature increases of greater than 2000°C accompanied by correspondingly high internal pressures are possible [5]. Fracture of the beta"-alumina electrolyte, then, is the most serious failure mode, and to prevent cell rupture, it becomes essential to limit the rise in cell temperature in case of electrolyte fracture [5].

The temperature rise due to the immediate reaction of sodium and sulfur after electrolyte fracture will be very rapid and governed by three factors: (1) the amount of reactants available, (2) the overall heat capacity of the battery system, and (3) the rate of heat dissipation [5]. Ofthese factors, only thefust, the amount of reactants, provides a useful way to control temperature rise. For example, to increase the heat capacity of the system sufficiently to limit the temperature increase generated by the production of 1 kg of Na2� to 200°C, almost 4 kg of stainless steel would be required. Furthennore, although heat can be dissipated to surrounding cells, heating of adjacent cells could also propagate cell failure [5]. Minimizing the amounts of sodium and sulfur that can react immediately becomes the first principle of cell safety.

The amount of bulk sodium available at the surface of the beta"-alumina electrolyte controls the rate of heatgenerated by the secondary reaction of sodium andsulfur upon electrolyte fracture. Ifthe minimum flow rate of sodium to the electrolyte, which is set by the maximum discharge rate under normal operating conditions, is not greatly exceeded, the heat should be dissipated over a reasonable period of time by the heat balance designed into the battery system [5]. Cell design to minimize the flow of sodium to the electrolyte surface is described below in the section on Na/S cell design.

Even if the increase in cell temperature upon electrolyte failure could be limited, failure propagation could occur if seals separating the reactants fail or if the outer cell containeris breached through corrosion by sodium polysulftdes at high a temperature [5]. Thus it is critical for cell safety to protect both the seals and the outer cell container against corrosion.

5 TP -4678

Seals

u The beta"-alumina electrolyte separates compartmentscontaining liquid sodium and sulfur/sodium I' polysulfides, highly reactive and corrosive materials that must be sealed from eachother andfrom the l atmosphere. The compartments must also be electrically isolated fromeach other. Electrical r (, conductivity anddurability largely determine the selection of materials andfabrication techniques for the li electrolyte;consequently, methods to seal the cell are limited becausethese methodsmust be compatible l with the materialproperties of the electrolyte [5]. Seals are a critical component of cell structure both for c reliabilityand safetyconsiderations, because they can limit the maximum temperature and pressure the r l cell can attain [5]. Ifthe seals fail, reactantscan leak out and breach the cell container, which in turncan L lead to electricalshort circuiting of other cells in the battery. r i Alpha-Aiumlna-Beta"-Aiumina Glass 'C,

Most cell designs use an alpha-alumina collar (also called header or top cap) to provide electricaland ionic insulation between the two metallic current collectors. Usually, the beta"-alumina electrolyte is bonded to the alpha-alumina by a glass sealant. Many configurations for this seal are used. The glass material must provide a hermetic seal between the header and the electrolyte andwithstand attackby liquid sodiumas well as an occasional freeze-thaw cycle. Ideally, the glass seal should be stress-free at the operating temperature of the cell and have a net radial compressive stress at room temperature [5]. The properties of glass used for seals should include [7]: ' I \, • Thermodynamic stability in Na, S, andsodium polysulfides • Low sealing temperature to minimize interdiffusion of electrolyte and glass components • Thermal expansion coefficient matching those of the electrolyte and alpha-alumina seal • Thermal stability at cell operating temperature • Ease of fabrication. r According to Sudworthand Tilley [5], an enormous amount of effort hasbeen devoted to developing this l seal, and the seal is "almost fully developed now in a satisfactoryform .... " Severalcells with a aluminoborosilicate glass seal used in all standard production cells made by CSPL have remained intact (as of 199 1) for more than8,800 electrical cycles in more than 62,000hours [8].

Metal-to-Alpha-Alumina

Early metal-to-alpha-alumina seals were mechanical compression devices using a variety of gasket materials and clamping methods [5]. Contemporary seals use a thermocompression bondingtechnique in which the simultaneousapplication of heat andpressure creates atomic interdiffusion of a metal and a ceramic. For example, in its PB cell design, CSPL has used a direct ceramic-to-metal bonded seal for the sodiumelectrode closure and a metal-aluminum interstrateseal for the sulfur electrode closure [9]. The upper temperature limit for present seals is about 650°C [5].

l,

r " !' 6 L

L TP -4678

Cell Container-sulfur Electrode

The third principle of cell safety, according to Sudworth and Tilley [5], is to protect the outer container from corrosionby sodium polysulfides at high temperatures. Ina sodium-core cell, the outer container houses the sulfurelectrode and also serves as the positive current collector. The formidable material requirements for the container-electrode include goodcorrosion resistance,high conductivity,ease of fabrication,low weight, and low cost; consequently, no single material that combines all of these properties hasbeen found [5]. Corrosion resistanceis required not only to maintain mechanicalintegrity of the cell, but also to prevent degradationof cell performancebecause corrosionproducts can increase cell resistance [5]. Most batterydevelopers have tried to meet these difficult material requirements by using compositematerials, usually an inexpensive substrate,such as aluminum,carbon steel, or stainless steel that is coated, plated, or sheathedwith at least one layer of corrosion-resistantmaterial [9].

The complexityof the electrochemicalreactions in the sulfurelectrode complicates an already difficult problem of material science. In the sulfur cell, the electrochemical potential,stoichiometry, and compositionof the polysulfide melt varyperiodically during cell operation andchange the solubility, wetting properties,chemical reactivity, and viscosityof the reaction products [5]. Although researchers have tried numerous combinations of metals, nonmetals,alloys, and ceramics to develop a suitable material for the sulfur container [5], it remains a major unsolved problem of material science for Na!S battery technology [10]. Current Na!S cell designs use a chromium-containing layer as the primary corrosion barrier [10]. Chromium also seems to play an important but yet unknown role in the recharging of Na!S cells; according to one expert, all the cells that work well use chromium for corrosion resistance [10]. A patent issued to Ford Motor Company [11] claims that continuous, controlled release of ions of chromium and other metals improves cell performance, reduces corrosion of corrodible metal current collectors, andprolongs cell life. A betterunderstanding of the role of chromium (and other transition­ metal sulfides)in the electrochemistryof the sulfurelectrode would also help in developing better corrosion-resistant materialsto contain the sulfurelectrode.

Na/S Cell Design

All threeprinciples of cell safetyhave been incorporated in the present generation of Na!Scells. The ABB Na!S celldesign hasnot changed very much since about 1985 as ABB has focusedits RD&D on batteryengineering and battery-vehicle interfacing [10]. The major components of the ABB cell area cylindrical metallic cell casing approximately 35 mm in diameter and 230 mm in length that also serves as the current collector, a sulfurelectrode of graphite felt impregnated with sulfur,a tubular beta"­ alumina electrolyte surrounding an internal metal container and sodium (Figure2). The cell is assembled by inserting the sulfurelectrode in the cell casing,followed by the electrolyte tube containing the sodium electrode. Aninert gas is used to force sodium onto the surface of the electrolyte at all states of charge. The electrolyte tube is closed offby a beta"-alumina-glass-alpha-alumina seal and an alpha­ alumina-metal thermocompression seal [3]. This cell design, designated A04by ABB, hasa capacity of 38 ampere-hours (Ah). Recently, ABB announced a A08 cell, which is similarin configuration to the A04cell but which hasa slightly higher capacity [13].

7 TP-4678

'i: l

c ,, l

Sodium r L Sulfur+ C-felt f �� AI casing \_

R

Electrolyte

Safety insert '

Figure 2. The ABB cell design Source: Ref. 24.

8

r TP-46 78

The CSPL cell design contains essentially the same components as theABB design but differsin geometry. The CSPL cell design, designated PB (Figure 3), is smaller thanthe ABB A04cell, with a capacity of 10 Ah, and is shorterand can-shaped rather than tubular in shape. By decreasing the capacity of the cell, CSPL's design isolates small amounts of sodium(1 4.5 g) and sulfur (2 5 g) at the cell level and, similar to the ABB cell, incorporates important safety features into the cell design [9]. Although the geometry of the PB cell has remained essentially unchanged, its internalconfiguration has evolved over several designgenerations, resulting in both improved performanceand safety. (See Molyneux et al. [14] for a history of CSPL's cell designs). To improve safety, internal design changes were made to limitthe quantity of sodium available for reaction with sulfurif the electrolytefractures, to restrict the flow of sodium to the site of the electrolyte fracture, andto protect the metal casing (and current collector)from corrosion by sodium polysulfidesat high temperatures [9]. The bulk sodium is stored in a corrosion­ resistantsafety tube, which has a small supply hole at its base. The safety tube fits within the beta"­ alumina electrolytetube so that the annular distance between the two tubes is very small [10]. A small hole in the safety tube keeps the flow of sodium as small as possible and yet allows the minimum amount of sodium needed to maintain the electrochemical reaction.

Cell Safety Testing

The principal mechanisms of Na/S cell failure include cracking of the electrolyte, degradation of the glass seal, corrosionof the container, and breaching of the metal-to-ceramic seals. Operating factors that impact these failure mechanisms include number of electrical and thermal cycles, time at operating temperature, depth of discharge, and cycling regime. Cells aretested for performancereliability and safety by subjecting them to electrical andthermal cycling. Cells aretested for both lifetimeperfo rmance under normaloperating conditions and for behavior under failure or extreme conditions.

Life Testing

Cells areelectrically cycled undera variety of conditions in the laboratory. Over the pastdecade, the standard test regime has been a 3-hour, constant-current or voltage discharge and a 5-hour charge. Cells have also been tested using a number of other test regimes, including simulated on-road cycles such as the FederalUrban Driving Schedule (FUDS) [ 4] anda simplifiedversion of FUDS called SFUDS. The standard test regime for electric vehicle (EV) applications is moving toward SFUDS.

There has been a steady improvement in results of service life testing at CSPL due to improvements over the past 5 years in the PB cell design. Test resultsshow that safety performance of the PB cell design has also improved considerably between 1985 and19 89, with no breaches of the cell container or the seals in more than2 yearsof life-cycle testing and abuse testing (described in the next section). SandiaNational Laboratories (SNL) reported in 1990 that CSPL has had no infant mortality in the last 2400PB cells, a failure rate of 0.0007%/cycle (measured to more than 1000 cycles) anda 0.0001 probability of a cell temperature exceeding 450° C [15]. Extensive life testing at the cell level has established the reliability of the PB cell design with virtual elimination of infant mortality on initial heating. The results of continuous testing for more than 3.5 years at CSPL to qualify cell components are summarized in Table 2. These results show that the major components of CSPL' s central sulfur "TD" cell arestable and have qualified out beyond 5 years [7).

9 TP-4678

i l

r l

r I ,,

\;r

J \

"' 8 � I.t ., 0

a>�

...... r ...... l ......

......

......

• • • • • 0 ••• f ...... l

Minimal amount of readily available sodium for reaction

Sodium flow restricted

Figure 3. The CSPL cell design Source: Ref. 15 .

10 TP-46 78

Failure Testing

During failure testing,cells aredeliberately damaged in order to assess their behavior under conditions far beyond normal operating conditions. The main failure testing method used at CSPL is to initiate cell failure by deliberately applying about 50 V to a fully charged cell [4]. This overvoltage leads to a catastrophic fracture of the ceramic electrolyte. Electric current is then passed through the cell to simulate failure in a battery environment, and the temperature exotherms generated by the directmixing of sodium andsulfur are monitored by thermocouples attached to the cell case[16 ]. After the reaction of sodiumand sulfur is completed and the cell is cooleddown, its condition is determined through post-test examination, including X-ray radiography [16]. More than100 failure tests of PB Mkill SF cells have been completed (as of April1989) at CSPL, and no breachinghas occurred [17]. Post-test examination by X-ray radiography showed disintegrationof the electrolyte and completereaction of the electrodes [17].

Failure tests at CSPL comparingPB cells with and without safety features (primarilythe metal safety tube containing and limiting the flow of sodium to the solid electrolyte, as described earlier) showed that these features moderateexcessive temperature exotherms and prevent subsequent breaching offailed cells [16]. These tests were performedusing 36 -cell submodules by failing cellssequentially and monitoring the temperature exotherms of these cells. Most of the cells without safety features failed with maximum temperatures below 400°C. The temperature in 3 of 23 cells tested exceeded800°C and causedcell breaching. Damage from the breachingcells was limited to one or two adjacent cells, and there was no propagation of failure through the submodule [16]. In contrast, more than30 tests on cells withsafety features showed a mean temperature exotherm of 42°C (from a starting temperature of 350°C), with a maximum temperature under450 °C. None of these cells sustained cell breaching [16].

CSPL has examined the failure characteristics of the PB Mkill cells that breachduring safety testing [16]. Once the temperaturein these cells reached about 500 °C, the risein temperature increased rapidly and caused breaching of the cell seal due to the combined effects of high-temperature corrosion andcell pressure. When the temperature of the cells test exceeded 525°C, the pressure of the sulfurelectrode becamegreater than that of the sodium electrodes, which weresealed at610 Torr. It is at this point that breaching of the cell container is initiated in cells without safety features. To explore the sensitivity of cell breaching to relative electrode pressure, CSPL constructed nine cellsin which sulfur pressure was higher than that of sodium at the normaloperating temperature of 350°C. When these cellswere failure tested at this temperature, seven of the nine cellswere breachedimmediately after fracture of the electrolyte, with accompanying large temperature exotherms. This experimentby CSPL showed the need to maintain the pressure of the sodiumelectrode abovethat of the sulfurelectrode. Furthermore,an increase in the startingpressure over eitherelectrode could cause large temperature exotherms after electrolyte failure. The experimentalso showed the need to reduce the as-assembled pressureover both electrodes as much as is feasible.

11 TP-4678

Table 2. Cell Component Stability

Component Cycles Hours Comments € \

Electrolyte composition 4,900 40,392 f I' Glass composition 4,900 40 ,3 92 (Type1 glass) 44,520 (Type2 glass)

Thennocompression seals 5,454 44,520

Sodium electrode 5,454 44,520

Sulfur electrode 5,454 44,520

Containment 27,500

Secondary protection of 5,454 44,520 containment (carbon coating)

r l

Source: Ref. 7.

' ' ' ,,

12 TP-4678

CSPL has also subjectedPB cells to elevated temperaturetesting by deliberately overheating the cells [16]. These testshelp to establish the limit for seal integrityand showed thatthe meantemperature at seal failure for the Mklllcell was 605°C, at which point the sulfur vapor pressure reached 5320 Torr. The PB cell can maintain cell integrity to temperatures in excess of 600°C, but the combined effects of polysulfide corrosion and sulfur electrode pressure generated on cell failure limits the maximum acceptable temperature to about 525°C [16 ].

Thermal (Freeze-Thaw) Cycling

The durability of Na/S cells under thermal, or freeze-thaw (Fn'), cycling is an important consideration for safety as well as reliabilitysince most cell failures due to F/1' involve fracturing of the ceramic electrolyte or breaking of seals [18]. These failure modes can cause leakage of reactantsand pose a potential problem for safety. A Na/S battery will besubjected to numerous F/1'cycles duringits expected5- to 10-yearlifetime due to, for example, low-demand periods for vehicle use or when the vehicle is inoperable for mechanicalreasons. Although research, testing, and mathematical modeling of stress under F/1'have greatly improved understandingof the problem, cell durabilityunder Fffremains a major area of concern [18].

The factors that influenceF/1' failure rate include cell design,cell lifetime, rateof heatingand cooling, and state-of-charge [18]. External factors such as shockand vibration could also influence F/1' failure rate [18]. Manyof the F/1'failures in earlier cell designs that had a high ratio of length to diameter (aspect) were encountered on initial heat-up. These infant mortalityfailures can be avoided through proper cell design, fabrication, and break-in procedures. For example, CSPL's change in cell design from a large (150 Ah) tubular design with a high aspect to the smaller PB design with a low aspect improved Fffperformance by reducing the mechanical stresses on the electrolyte due to a cantilever effect. Also, some alpha-alumina-to-metalseal designs mechanically decouple the electrolyte from the container [18]. As mentioned in the previous section, improvements in cell design have virtually eliminated infant mortality on initial heating in the PB cell [15].

The more difficult F/1'failures to prevent are those that occur after initial heat-up and breaking-in of the cell [18]. There has been extensive F/1'testing on individual cells to determine the effectsof depth-of­ discharge and thermal cooling/heating rate on solidification/melting of molten cell materials, the effects of multiple F/1'cycles on component integrity and electrical performance of relatively new cells, the effect of one Fffcycle on the electrical performance of cells with many electrical cycles, and the effect of rapid cooling and heating rates on cell integrity [18]. These tests showed that PB cells possess "excellent intrinsic F/1'durability at the cooling and heating rates that will be associated with an operationalbattery" [18]. Rapid heating and quench cooling ofPB cells with subsequent electrical cycling showed that performanceshould not deteriorate as a result of thermalcycling [18]. There are much fewer data available in the literature on the F/1'durability of ABB cells [12].

A major question remaining is the F/1'durability of modules and batteries. The relatively few modules that have been subjected to F/1' testing show a higher and sometimes unacceptable cell failure rate than that observed for single-cell F/1'testing [18]. Inone case, post-test analysis showed many types of cell failure, including seal and container leakage and electrolyte fracture [18]. The lackof correlation between F/1'testing results for individual cells and cells in modules raised questions concerning the F/1' durability of the PB cell as well as test procedures used to determine F/1'durability [19]. Since then it

13 TP-4678 was detennined that test procedures thatdid not allow the temperature to go low and long enough for freezing to take place were at fault [19]. Significantloss of battery capacity after a few Fffcycles is another indication that Fff durability may remain as an importantproblem [18].

Research and testing on Fff durability seem to indicate that Fff cycling accentuates existing failure mechanisms and thus accelerates cell failure; it is no longer thought thatFff cycling by itself causes r electrolyte failure in PB cells [12]. There is still uncertainty on what critical parameter of cell history I allows Fff cycling to accelerate cell failure. Possible parameters include time at temperature, number of l previous Fffcycles, number of coulombs passed, and contamination of the electrolyte surface [18]. Other important questions include whether Fff failure generally occurs during freezing or thawing and the importance of heating and cooling rates nearthe melting point of the positive electrode materials [18]. Better understanding of how Fffcycling accelerates cell failure mechanisms may be important to improved cell safety. r l

Safety Status of Na/S Cells

It is evident that RD&D over the past decade on cell design and component materials have improved the reliability and durability of CSPL' s PB cell to the point where cell safety per se is no longer a major concern under nonnal operating conditions. The major safety issue remaining for the PB cell is quality control as scale of production is increased. The Chloride-RWE pilot plant at CliftonJunction, UK, should begin to provide data on quality control of the PB cell in a production versus a laboratory environment. Unifonnity of cell aging is a safety issue at the battery level [10] and is discussed below.

There is considerably less infonnation on safety testing of the ABB cell. ABB seems to have focused its efforts on safety testing at the battery level. These safety tests are discussed below. Whether the safety established for individual cells in a laboratory environment is valid in a battery environment remains a critical question for further RD&D.

Battery Safety r i l_ Battery Design

A Na/S battery consists of many hundreds of mechanically and electrically interconnected cells in an insulated enclosure that also houses a thennal management system for initial heat-up and waste heat control. The safe use of Na/S batteries for vehicular propulsion begins with cell safety. Intensive RD&D : and testing have improved the reliability and durability of Na/S cells to the point where CSPL has enough confidence in the technology to begin pilot-plant production of the PB cell design and where ABB felt it could concentrate on battery engineering using its A04 cell design. Although cell safety has been established through life and failure testing, the electrical, thennal, and mechanical environment at the battery level imposes additional safety questions and requirements. The fundamental premise of cell safety, isolating sodium and sulfur in small amounts and controllingthe flow of sodium to the sulfur electrode, means that a large number of cells need to be interconnected to derive sufficient energy and power levels required for a battery.

r 14 TP-4678

The safety strategyfor batteries usedat CSPL is to build in layersof protection from thecell up [20]. CSPL has constructedand tested a full-size battery containing 3584cells in foursimilar quarter-unit modules [14]. The batterycapacity was 300 Ah at 230 V, and it weighed850 kg [4]. In the CSPL battery design, cells are assembled in banks that arewrapped in an electrically insulating blanket. These banks are assembled into quarter-battery modules, which in tum are placed in a thin, evacuated, double-wall, steel outer container. Each level of protectionis designed to address the critical safetyissue at that level of battery assembly-from double-wall containment of sodium at the cell level, to electrical insulation to protect against short-circuits at the bank level, to thermal insulation and double-wall containment at the outermost level.

The important components of battery design areelectrical networking, electrical insulation, and thermal management [14]. A goodelectrical network design will accommode cell failure with minimal effect on batteryperformance. CSPL based its battery design on a four-cell string subunit with parallel interconnectionsat the ends of each string [14]. If a cell fails under such an arrangement, the remaining three cells in the string will be charged by parallelstrings. One of the three cells will then develop a high effective resistance characteristic of Na/S cells at top of charge because a cell in which sodium ions have been depleted from the sulfur electrode will no longer carry current in that direction [4]. Thecell then acts as an.open circuit device and prevents further currentflow into the string. The string is inoperable, but other strings in parallel with it should continue to operatenormally [14]. Each batterybank is individually monitored electronicallyto detectcharge imbalance caused by cell failure or cell aging. Fully charged banks can be by-passed and current then flows at a reduced level to the other banks [4]. The monitoring system is interfacedwith the charger and the vehicle controller to limit the maximum and minimum voltage across each bank [4].

The CSPL.battery design using a largenumber of low-capacity cells provides a multitude of parallelpaths and minimizes the effect of a cell failure on battery capacity. The effect of cell failure on battery safety is also minimized by the use of small, low-capacity cells because the amountof reactants that can escape froma failed cell is also small. The use of a smallnumber of cells in a stringalso reduces the overvoltage (which is proportionalto the number of cells in a string)experienced by the first cell to be recharged in the string[ 14] and thus reduces the likelihood of electrolytefracture. Battery design must also minimize the likelihood of failure propagation by providing proper thermaland electrical insulation. For example, CSPL removed aluminum finsused in an earlier battery design for passive heatdissipation because of the potential for a high-temperature excursion to create alternate conduction paths for shortcircuits and to further propagate cell failures [12].

The thermal management system, the third major component of Na/S batterydesign for safety, provides initial heat-up and controls waste heat while insulating the battery to maintain internal operating temperature and minimizing thermal hazard to the external environment. During normal operation, internal resistance losses will maintain operating temperature without additional heating [4]. The system must also maintain temperature uniformity within the battery enclosure because temperature gradients will affect cell performance. A typical temperature gradient between the inner andouter cells in prototype batteries is about 20°C, which does not present a significant safety issue [12]. Because the temperatureof the batterywill rise with the rate of discharge, the rate of heat removal is an important consideration in keeping the battery within temperature specifications [21]. In some CSPL battery designs, a forced aircooling system was activated when the internalbattery temperature reached about

15 TP-46 78

360°C [22]. If the temperature continues to rise, an electronic monitoring andcontrol system signals the controller, and a reduction in power is initiated [22].

The outer thermal container keeps heat loss within an acceptable range, insulates the battery from external temperature extremes, and forms the final protective barrier between any leaking reactants and the external physical environment. The thin double-wall evacuated enclosure must withstand water andsalt spray, mechanical shockand vibration, and impacts with road debris[4 ]. The vacuum insulation system r must also be load bearing and should have a heat conductivity of 1Q-3 W/m°C or lower [3]. The thermal '- conductivity of currentvacuum insulating systems is about0. 004W/m °C, andtypical thermal loss from a 50 -kWh battery will be about 200 W. This means that a battery can stand off-charge for up to 48 hours r [4]. Better vacuum insulation enclosurescould improve the safety of Na/S batteriesby potentially l reducing the number and severity of Fff cycling.

The ABB battery that has undergone themost extensive testing is designated as the B-11 . The B-11 consists of 360A04 cells, with the most common configurationbeing six 60-cell stringswith cross­ connections every five cells [21]. The B-11 battery is an experimental battery designed for performance [ tests in EV applications and, depending on the cell configuration and electrical interconnectionpattern, \ has a continuous powerrating of 24-35 kW and weighs 200 -2 80 kg. Two B-11s would be used to propel a mini-van whereas three would be used for a full-size van. The battery uses replaceablecells and a conventional thermal plug to allow maintenance. Commercial versions of a smaller 10 -kWh design (B- 120 ) will be simplifiedand have nonremovable cells, complete vacuum insulation,and a shape tailored for EV applications [21].

Battery Failure Testing

Two ABB prototype batteries, the B-11 and the B-06, have beensafety-tested by ABB in cooperation with the German Technical SupervisingAgency (TechnischerUeberwachungs- Verein, or TUV)and, based on these tests, ABB has obtained permission to operateEVs powered by Na/S batteries on public streets in (then) West Germany without restrictions [21]. Similar approval has been obtained in Canada [1]. Data from the ABB-TUV tests are not available in the literature, but the test results show that the prototype batteries assembled by ABB based on the A0 4 cell design can meet German and international f standards(DIN, ISO) as well as additional standards established by the TUV for safety in vehicular use. I l It should be noted that these safety tests were performed on the batteries themselves and not on vehicles equipped with Na/S batteries.

The testing criteria for the TUV certificationare listedin Tables 4 and 5. Fifteen of ABB' s B06 and B11 Na/Sprototype batteries were tested against the following criteria [2 3, 24]. ( o Vehicular crash simulation (sudden deceleration) -a B-11 battery was mounted on a carrier that was accelerated to about 50 km/h and stopped in about 50 em. The resulting curve of velocity versus acceleration showed a maximum deceleration of 24 G and a mean value of about 18 G for 80 milliseconds, which is comparable to results in standard crashtests for passenger cars. Electrical cycling of the battery before and after the test showed no change in performance.

r 16 l TP-46 78

o Deformation testing-the thermal enclosure of a B-06 battery was compressed 25% to induce breakdown of cells. The battery was not destroyed, and there was no release of active materials. However, many cells failed and an internalfire occurred.

o Electrical short-circuiting-both B-11 and B-06batteries were also short-circuited by connecting the battery plugs by a lead with a resistanceof about2 milli-ohms. The currents increased to very high values (1 200 to 1500 amps) but stopped because the internal connectors melted after a few seconds. The cells remained tight, and the temperature increase was within the acceptable range.

o Effects of a vehicle fire-a gasoline fire was set for 10 minutes underneath a B-06 battery. The change in temperature within the thermal enclosure was less than10 K [3].

o Thermal shock-a hot battery was immersed into cold water. Only small leaks occurredin a few cells.

Based on the results of these tests, ABB maintains that the only safety issue that remains to be resolved is mishandling of the battery, for instance, overcharging with a high overvoltage [24].

Battery Life-Testing

Failure testingof prototype Na/S batteries has shown that they can withstand extreme conditions to which they may be exposed. Sodium-sulfur batteries must also undergo life-testing to be shown to be safe under normal operating conditions, which requires long-termfield testing under actual driving conditions. The first Na/S battery for an EV was constructed in England in 1975 by the Electricity Research Council Centre and�onsisted of 96 0 interconnected 30-Ah, central- sodium cells [5]. The battery was installed on a Bedford van, and the van was driven on public roads to demonstrate that a Na/S batterycould power a vehicle. The van had a range of 60 to 150 miles, but the batteryfailed after only 20 cycles [5]. A number of other one-time applications of Na/S to power vehicles are described by Sudworthand Tilley [5]. Since 1986 , ABB batteries have been installed in 80 vehicles, and20 of these vehicles powered by a B-11 battery, including a BMW, a Daimler-Benz190 , a VWJetta, and a Chrysler mini-van, have accumulated more than25 0,000km of operation (as of July 1990)in urban conditions on highways [19].

InJanuary 19 89, CSPL delivered a 19-kWh, one-thirdbattery module to Argonne NationalLaboratory (ANL) for testing. This module contained 960 PB cells and included a charger and a thermal management system [21]. The module was evaluated for 241 test cycles by ANLand performed satisfactorily over this number of cycles. The thermal management system maintained the battery within the operating temperature range of 330° -3 60°C throughout the testing. The module showed a high initial discharge battery resistance after a full recharge,which limits available power for accelerationand can cause high charge voltages if regenerative braking currents arenot limited [2 5].

17 TP-4678

Table 3. Program for Safety Tests on Sodium-Sulfur Batteries

Part 1: Normal Use

Item Criteria of Acceptance l

Electrical insulation to ground R>500 0hmN f: i il High voltage Protection should be guaranteed by battery control unit High current r State of charge Shown at the panel

External short Special insulation of current leads Fuse

Surface temperature of thermal enclosure Maximum values are at specified points with/without leaks atthe vacuum insulation on the outer battery casing

High temperature Protection should be guaranteed by battery control unit and/or designof heaters

Vibrations (DINIEC 68 ) No ground faults • 2 G vertical No cell failures • 1 G horizontal (both axes) • 10 min at resonance frequency • 5 h discharge current

Source: Ref. 23.

18 TP-4678

Table 4. Program for Safety Tests on Sodium-Sulfur Batteries

Part 2: Safety During/After Accidents

Item Criteria of Acceptance

Mechanical shock (crash) No damage of battery or cells (20 G, 70 milliseconds)

Deformation No external fire (up to 25 percent, breakage Emission of sulfur dioxide acceptable of cells)

External short No externalfire (bridging of plugs) No emission of sulfur dioxide or other reactants

Upside down Batteryoperation is not affected

Source: Ref. 23

19 TP-46 78

In May 1990 , ANLreceived an 8-V, 300-Ah module from CSPL for both performanceand life testing

[26]. The module contained 120 cells configured into 30 parallel-connected stringsof 4-series connected t cells. in ! The module uses the same cell design and hardwareassembly used the full-sizedbattery system li for the Ford ETX-11 vehicle described below. Aftercompleting perform ance testing,ANL began life ,_ testing in October 1990. As of July 1991, the module has completed more than500 [charge-discharge] cycles, while retaining about 91 % of its initial capacity of 292 Ah(at a 3-hrrate).

InJanuary 19 90 , CSPL delivered a full-sized battery composed of three19 -kWhenclosed modules for testing in the experimentalFord ETX-11 vehicle. A groundfault (laterdetermined to be caused by f shipping damageand poor placementof heaterwire leads) was detected afterthe battery was uncrated L [19]. Limited testing of the battery before it was returned to CSPL showed that the battery met the " performance and physical requirements of the contract under which it was made [19]. The battery was I repaired by CSPL, who also modifiedthe thermal, electrical, mechanical, and electronicsystems of the l batterybased on the results of the limited testing and evaluation of damage from shipping [15]. The battery was re-delivered to Ford in August 1990 . The battery is now being tested at Idaho National Engineering Laboratory and has been at temperatureand cycling for nearly 18 months [27].

In May 1990 , ANLreceived a B-11 battery (90 V.22 kWh) from ABB for performance characterization and life testing [26]. The battery has 360 cells (30 Ah each) in three series-connected sections,with each sectioncontaining eight parallel-connected strings of IS series-connected cells. Life testing began after the batteryhad been operatedfor about 16 0 cycles to characterize performance. Afteran interruption in testing to repair an electrical connection and to disconnecta cell (and correspondingcells in each of the stringsin that section) that had failed upon cooling (for the repair) andheating back to operating temperature, ANL resumed life testing; as of July 1991 , the batteryhas retained about 89% of its initial r capacity of 238 Ah after more than450 cycles. Moreover, the thermal management system of the battery l has maintained the battery temperature withinthe operatingtemperature range (3 30°-350°C) for high­ rate discharges to 50 W/kg.

\.._

Safety Status of Na/S Batteries

Prototype Na/S modules and batteries have been tested under laboratoryconditions for both performance and safety. Laboratory testing has, not surprisingly, been focused on performancesince prototype batteries are few in number, expensive, and are tested to meet contractual specifications based on performance standards. Safety testing has been limited primarily to those conducted by ABB in l qualifying the B-06 and B-11 battery prototypes for use in test vehicles in Germany. Batteries have not been safety tested in vehicles. Field trials of test vehicles have provided dataand information on performance and evidence that vehicles powered by Na/S batteries canbe operated safely in urban driving conditions. Although a significant number of kilometers have been accumulated by vehicles using ABB batteries, these vehicles are experimental and, therefore, limited in number and driven under controlled conditions.

r 20 l TP -4678

Conclusions and Recommendations

Both failure testing andlif e-testingof cells have shown thatthe chemica l andthermal risks implicit in the presence of sodium, sulfur, and sodium polysulfideshave been satisfactorilymitiga ted by engineering anddesign at the cell level. These modesof testing, however, need to be combined to evaluate the hazards and risks posedby Na/S batteries fr om cell failure over the lifetime of the battery. Predictionsof batterylif e based on cell statistics have not been accuratebeca use of thediff erencesbetween bench and field testsand the difference between cycle life and calendarlif e. Cell testing is useful fo r ranking the effectsof celldes ign changes; understanding of batterylif e requirestes tingof full-sized batteries [19].

Cell failure statisticsderived from laboratoryte sting procedures show that failure rates, particularlythose due to infant mortality, have been reduced. There is still a finiteprobability of cell failure, however, even under a laboratorytesting environment fo r individual cells. Since attaining a zero fa ilure rateis unlikely, especially at the batterylevel under actual operating conditions, the effectof infrequent, isolated cell failures on batterysaf etyis an open question [12]. A failedcell in a batteryenvironment is su sceptible to corrosion and breaching of the cell container, particularly over the 5- 10 yearlif etime projected fo r commercial Na/S batteries.

Commercial Na/S batterieswill be enclosed in a vacuum insulating outer container and will, in effect, be maintenance-free. Individual cells cannot be tested andremoved as they can be, fo r example, in ABB's B- 11 prototype battery. Furthermore, spatial distribution of cell failures has a significant effect on battery performance as well as undeterminable consequences fo r battery safety. In other words, individual cells in a battery'sh ould not only have uniform propertiesto begin with, but they should also age at the same rate [10]. The safety of Na/S batteriesdepends on the electrical and thermal functioning of thousandsof individualce lls and the electrical, thermal, and mechanical sy stemsthat control and house these cells. A comprehensive, long-term testing program of Na/S batteries in vehicles under actual driving conditionsis needed to demonstrate that the technologyhas evolved to a point where commercialization is possible.

Based on the discussion above, the fo llowing conclusions andrecommenda tions can bemade concerning the sa fetyof Na/S batteries.

1. The ABB testsan d TUV approval based on these tests are not sufficient fo r U.S . Federal Motor Vehicle Safety Standards(FMV SS) acceptance fo r the fo llowing reasons:

• Only prototype batteries were tested

• Batterieswere tested outside of a vehicle

• Batterysa fety tests consistedof simulations of a vehicle crashing, catching fire, and being submerged in water

• Testing results but not data have beenrelease d by ABB andTUV .

21 TP-4678

2. The ABBII'UV tests may be more useful for obtaining shipping regulations thanfor obtaining FMVSS certification.

3. The tests performed by ABB for TUV approval (deceleration, deformation, fire, thermal shock, electrical short-circuiting)should form a sufficientbasis for safety testing of the battery for catastrophic accidents.

4. It is not possible at this time to determinethe type, number, and test regimes needed to meet the FMVSS for Na/S-poweredEVs, especially crashworthiness. f 5. The DOE andDOT should work with industry, perhaps through the Vehicle Safety Sub-working l Group of the Ad Hoc EV Battery ReadinessWorking Group, to develop a comprehensive test plan for Na/SEV s to meet, at a minimum, the applicable FMVSS requirements.

6. The number of batteries required for destructive testing both outside and in vehicles must be determined after the testing plan is developed, not before.

7. Vehicle safety cannot be considered apart from design requirementsfor EV s. One example is the impactof new glazing materials, such as electrochromicglass for improved heatingand cooling efficiency, on the ability of EVs to meet the requirements of FMVSS No. 205 on glazing materials.

8. The issue of attaining a public perception of the safetyof Na/SEV s will probably be very different from that of obtaining FMVSScertification. Battery testing outsideof the vehicle may be very important, since public perception may focus on the most obvious difference between EVs and internal combustion vehicles. Long-term, day-to.,-dayoperat ing convenience and reliability will also affectthe public's perception of safety. Comprehensive and widespread public information and l demonstration. programs will be critical.

9. Destructive testing andcrash testing do not address the other importantquestion of long-term safety under normal operating conditions. Questions about cell aging and the effectsof isolated cell failures within a battery must be addressed in a long-termtesting program.

�-

: '-

r

22 l TP-46 78

References

1. P.O. Patilet al., "Shipping, Use, and Disposal/Recycle Considerations for Sodium/Beta Batteriesin EV Applications," Proceedings of the BetaBattery Workshop, Chester, England, June 12 -14, 1990.

2. JJ. Cohrssen and V.T. Covello, Risk Analysis: A Guide to Principles andMethods fo r Analyzing Health andEnvironmental Risks , Washington DC: Council on Environmental Quality,19 89.

3. W. Fischer, "SodiumSulfur Battery Accomplishments and Remaining Problems," presentationto MaterialsResearch Society Conference, Boston, Nov. 28-Dec.2, 1988.

4. M.F. Mangan, "SodiumSulphur Batteries for ElectricRoad Vehicles," Proceedings, EVS9, Toronto, Nov.13-18, 1988.

5. J.L. Sudworth andA.R. Tilley, The Sodium Sulfur Battery, London: Chapman and Hall,19 85 .

6. N.I. Sax andRJ. Lewis, Dangerous Properties of Hazardous Materials, 7th ed., New York: Van Nostrand Reinhold, 1989.

7. F.M. Stackpooland S.MacLachlan, "CorrosionResistant Materials for Sodium SulfurCells," Proceedings: DOEIEPRI Beta (Sodium/Sulfur) Battery Wo rkshop VII , Oct.1988, 24-1 ff.

8. N.J. Magnani et al., Exploratory Battery Technology Development Reportfo r FY90, SAND91 -0672 , April '1991 .

9. N.J. Magnani et al., Exploratory Battery Technology Development and Testing Reportfo r 1988, SAND89-3039, Oct.1989.

10 . J. Rasmussen, Beta Power, Inc., personalcommunication, Oct.11 , 1990 .

11 . U.S. Patent 4,76 7,684 (Aug. 30 , 1988), "Method of ChargeabilityImprovement and Reduced Corrosion for Sodium-sulfurCells," assigned to CarlR. Halbach. am I indebted to J. Rassmussen and J.W. Braithwaite for this reference.

12 . J.W. Braithwaite, Sandia National Laboratories, personalcommunication, Oct. 4, 1990.

13. W. Fischer, "Status and Prospects of Sodium-sulfurHigh Energy Batteries," Proceedings, 26 th IECEC, vol. 6, 80 -87, Aug. 1991 .

14. J. Molyneux et al., "Development and Testing of Sodium Sulfur Batteries for Electric Vehicle Applications," Proceedings, 22nd IECEC, 975 -982 , Philadelphia,PA, Aug. 1987 .

23 TP-4678

15. SandiaNational Laboratory, Presentation at Electric Vehicle Program Laboratory ProjectReview Meeting,Argonne NationalLaboratory, Sept. 6-7, 1990.

f I 16. F.M. Stackpool, "Cell Development-A Focuson Safety," Proceedings: DOF/EPRI Beta II t (Sodium/Sulfur) Battery Wo rkshop VII, Oct.1988.

17. F.M. Stackpoolet al., "Sodium Sulphur Battery Development," Proceedings, IEEE 1989,2765- f 2768.

. 18. J.W. Braithwaite et al., "Effect of ThermalCycling on CSPL Sodium/SulfurCells," Proceedings: f t DOF/EPRI Beta (Sodium/Sulfur) Battery Workshop VII, Oct.1988, 38-1 ff. '

19. J.W. Braithwaite, SandiaNational Laboratories,"ForeignTravel Trip Report Summary," July 12, 1990 (unpublished).

20 . P. Binden, Beta Power, Inc., personal communication, Nov. 26, 1990.

21 . M. Altmejd et al., "Sodium/Sulfur Batteries for EV Propulsion," Proceedings, IEEE,1989, 2769- 2774.

22 . R.D. McDowall, Safe tyIs sues Associated with the Testing of Sodium Sulfur Batteries, Internal Technical Report,PG-EHVP-90-064, EG&G Idaho, Inc.

23. W. Dorrscheidt et al., "Safety of Beta Batteries," Proceedings: DOF/EPRI Beta (Sodium/Sulfur) Battery Workshop VII , Toronto, June 1-3,19 88,46-1 ff.

24. W. Fischer and T. Shiota, "Stateof Development of Sodium Sulphur Traction Batteries at ABB and Powerplex," Proceedings, EVS 9, Toronto, Canada, Nov. 13-16, 1988.

25. W.H. DeLuca et al., "PerformanceEvaluation of AdvancedBattery Technologies for ElectricVehicle Applications," Proceedings,25 th lntersociety Energy Conversion Engineering Conference, Reno, NV, Aug. 12 -17, 1990, PaperNo. 900051.

26. W.H. DeLuca et al., "Performance and Life Evaluation of Advanced BatteryTechnologies for ElectricVehicle Applications," SAE Technical Paper Series 911634, reprinted fromAlternative Fuels in the Nineties (SP-87 6), Future Transportation Technology Conference and Exposition, Portland, OR,Aug. 5-7, 1991.

27.A.F. Burke, Laboratory Testing of the CSPL Sodium-sulfur Traction Batteryfo r the ETX-II Vehicle, Report No. EGG-EP-96-88, Dec.19 91.

F

I t

r

24 t

'

L Appendix A

Chemical Hazards of Sodium-Sulfur Batteries

Source: Sandia National Laboratories Safe Operating Procedures fo r Sodium/Sulfur Battery Laboratories July 11, 1990

I. Sodium

A. Sodium - Its Nature

Sodium (Na) is a soft, lightweight metal easily cut with a knife. It is so reactive that it does not occur in elemental fonn in nature but is widely distributed in stable compounds. In an inert atmosphere, freshly cut sodium has a pinkish bright metallic luster. In air, the cut surface fonns a white or grey coating of sodium oxide, hydroxide, and/or carbonate. Nitrogen, argon, and helium are inert to sodium and can be used to backfill containers used to store sodium. Hydrocarbons, such as mineral oil, kerosene, xylene, toluene, etc., neither dissolve sodium nor react with it, and are commonly used as storage media for finely divided sodium particulates. Sodium metal is a good conductor of heat and electricity. Sodium requires special handling techniques to avoid fire, explosion, and personal injuries.

B. Sodium Hazards

1. General

Sodium is a powerful reducing agent It reacts violentlywith water, forming sodium hydroxide (highly caustic and can cause serious burns) and hydrogen gas with release of considerable heat. The heat of the reaction with water is so great that in air or oxygen the released hydrogen invariably ignites with explosive violence.

2. Human exposure

Sodium causes severe thennal and alkali burns to body tissue, particularly to moist skin.

3. Other reactants to avoid:

- halogenated hydrocarbons (organic compounds containing bromine, chlorine, fluorine, and iodine: e.g., freon and trichloroethylene). - many metal halides (e.g., Ti, P) - hydrogen sulfide gas - sulfur dioxide - carbon dioxide at high temperature

A-1 - sodaacid or chlorinated fire extinguisher agents - mineral acids - metal oxides

The rates of the reactions cited in #3 can vary widely with the particle size of the sodium l and the temperature. The smaller the sodium particle and/or the higher the prevailing temperature the more rapid (ormore violent) the reaction. n. Sulfur

A. Sulfur - Its Nature I \i_ Sulfur(s) occurs both in the elemental state and in compounds, mainly as sulfides and r sulfates; it constitutes 0.05% of the crust of the earth. Elemental sulfur is an odorless, I tasteless, non-metallic, pale-yellow solid. Sulfuris insoluble inwater, sparingly soluble in alcohol andether, and soluble in carbon disulfide.

B. Sulfur Hazards

1. Human exposure

Sulfur dust is an irritant: inhalation of large amounts may cause inflammation of the nasal passages and mucous membrane.

Acute exposureto the skin may cause irritation, redness, and pain.

Acute exposureto the eyes causes irritation, redness, and pain. l

2. Explosion/ignition: f

Sulfur, especially as dust, ignites in air above 261

3. Other reactants to avoid:

- Aluminum (powder) -Ammonia - Bromates - Activated carbon - Metal powder (e.g., copper, iron, zinc) - Hydrocarbons - Strong oxidants (e.g., peroxides, chlorates, perchlorates, nitrates, permanganates

A-2

l_

L III. SodiumTetrasulfide

A. The Nature of Sodium Tetrasulfide

Sodiumtetrasulfide occurs as glass-like, dark red pellets or fine yellow powder normally having a weak odor of hydrogen sulfide. It is strongly hygroscopic and strongly alkaline. Its melting pointis 285oC and its decomposition temperature is greater than300 oC. It is chemically stableat roomtemperature. It is extremely soluble in water and moderately soluble in ethanol.

B. Hazards of Sodium Tetrasulfide

1. General

- releases toxic sulfur dioxide upon burning. - extremely corrosiveif molten

2. Human exposure

- breathing dust causes severe burns of respiratory organs.

- aqueous solutions, vapors, and mists are extreme irritants to the eyes and skin; briefcontact can resultin severe damage.

- swallowing the liquid irritates the mouth, throat, and tissues of the gastro­ intestinal tract. If swallowed, this chemical will react with stomach acids to liberate toxic hydrogen sulfide gas.

3. Other reactants to avoid:

- mineral acids (can generate toxic, flammable hydrogen sulfide gas) - oxidizing materials

IV. Sodium SulfurCells and Batteries

A. Chemical

1. The amount of sodium (Na) present in a cell ranges from 15 g to 90 g. It is contained in a sealed stainless steel can witha ceramic header.

2. Under normal testing, contact with chemicals is extremely unlikely.

A-3 3. In the event of a fireor accident, personnel canbe exposed to sodium monoxide (Na:z()), sodium hydroxide (NaOH), sodium(Na) , and sulfurdioxide (SOz).

a. Na:z(), NaOH, and Na arehighly alkaline materials that cancause irritation and rapid tissue destruction through chemical and thermal burns.

b. S02 contacting the mucous membranes will cause formation of sulfuric acid resulting in severe tissue burns. Edema of lungs and respiratory paralysis may fll occurwith high level exposures.

f 4. The following Permissible Exposure Limits (PEL's) have been established by OSHA: l

a. SodiumHydroxide (NaOH), 2.0 mglm3 b. SulfurDioxide (S{}z), 5.0 ppm.

f I· l

r I l

A-4 f r !i L Appendix B

Toxicities and Reactivities of Chemical Elements and Compounds Potentially Associated with the use of Na/S Batteries

Chromium (Cr)

Chromium is poisonous to humans in low concentrations. Reported lethal doses have been as low as 71 mg per kg of bodyweight Itis also an experimental tumorigen andsuspected carcinogen. Chromium is moderately reactive. Chromium powder will explode spontaneously in air.

Hydrogen Sulfide (BzS)

Very poisonous gas in low concentrations. Doses as low as 5.7 mg per kg of body weight have been known to be lethal. Concentrations of 800ppm have been lethal in 5 minutes. Lower concentrations are also lethal after a somewhat longer exposure. Hydrogen sulfide is anexplosive gas, and is very dangerous when exposed to heat or flame. It ignites on contact with many substances. When heated to decomposition it emits SO,..

Sodium (Na)

Although specifictoxicity data areunavailable, it is fairly clearthat ingestion of metallic sodium would cause severe burns. Metallic sodium is extremely reactive. It reacts exothermally, and frequently explosively, with water to yield hydrogen gas and sodium hydroxide (c.f.). Itcan reactviolently with a host of elements and compounds. When heated in air, sodium emits the toxic fumes Na20.

Sodium Hydroxide (NaOH)

Laboratory experiments have demonstrated that sodium hydroxide is lethal upon ingestion in quantities of approximately 500 mg per kg ofbody weight. NaOH is a severe eye irritant, even in very low concentrations (e.g. 1% solution). It is corrosive to skin, eyes and mucous membranes. Inhalation or ingestion of mists, vapors or dusts cause mentioned effects. Scarring and damage to the upper respiratory tract can occur upon inhalation. Sodium hydroxide is a strong base and is highly reactive. It can react violently with a host of elements and compounds, in some cases forming explosive compounds.

Sodium Monoxide (Na20)

This gas is moderately toxic upon inhalation or ingestion. It is corrosive to skin, eyes, and mucous membranes. In concentrations as low as 10 ppm it has demonstrated toxic effects. It is immediately lethal in concentrations above 1000 ppm. This substance is fairly unreactive, but can react viQlently with water.

B-1 Sodium Sulfide(NazS)

Sodium sulfide is mildly poisonous. It is highly corrosiveto human tissue. When exposed to heat or flame, sodium sulfide is flammable. It reacts violently with water, and acts corrosively upon many metals. r Sulfur (S) l

Elemental sulfur in solid form is virtually non-toxic and can be taken internally without injury. Inhalation of sulfur dust can irritate the eyes and mucous membranes of the respiratory passages. Sulfur presents a fire and explosion hazard because it has a low ignition temperature and a tendency to develop static charges. Sulfur become liquid at about 120°C. Liquid sulfur has a low ignition point and can produce severe burns. The presence of hydrogen sulfide is also a possiblity when handling liquid sulfur. When heated to decomposition, sulfur forms the toxic gas SQ.. When burned in air, sulfur forms the toxic gas SOz.

Sulfur Dioxide (SOz)

This gas is poisonous by ingestion and inhalation. It is immediately life threatening in concentration of 400-500 ppm, but concentrations as low as 50 ppm areconsidered to be toxic. This gas is also an experimental tumorigen and teratogen. Sulfur dioxide is moderately reactive, and forms sulfuric acid in the presence ofwater.

rr li Sources: N.I. Sax and RJ. Lewis, Dangerous Properties of Hazardous Materials, 7th ed., New York: Van l Nostrand Reinhold, 1989; Encyclopedia of Chemical Technology, vol. 22, 3rd ed., New York: John Wiley & Sons, 1983.

L

B-2